Table 1 Statistics of covariates used as input to the Machine Learning model GBDT. Statistics are calculated separately for the internal and external cohorts. For the ECG interpretations, type \(\{0,1\}^{11}\) indicates a binary vector. The position corresponds to the ECG lead used for the interpretation.
Characteristic | Type | Internal(n=1756) | External(n=1127) |
|---|---|---|---|
Age | Numerical | \(61(\pm 31)\) | \(60(\pm 31)\) |
Gender(male) | Binary | 936(53%) | 629(55%) |
Medical history | |||
Hypercholesterolemia | Binary | 693(39%) | 485(43%) |
Hypertension | Binary | 943(53%) | 803(71%) |
Current Smoker | Binary | 368(20%) | 283(25%) |
Diabetes | Binary | 509(28%) | 354(31%) |
Prior MI | Binary | 303(17%) | 245(21%) |
Angina | Binary | 42(2%) | 80(7%) |
Prior CABG | Binary | 166(9%) | 180(15%) |
Prior PCI | Binary | 124(7%) | 6(<1%) |
CAD | Binary | 349(19%) | 271(24%) |
Family history of CV disease | Binary | 204(11%) | 81(7%) |
Symptoms | |||
Other | Binary | 1753(99%) | 1124(99%) |
Chestpain | Binary | 992(56%) | 644(57%) |
Syncope | Binary | 103(5%) | 69(6%) |
Shortness of breath | Binary | 417(23%) | 282(25%) |
Diaphoresis | Binary | 114(6%) | 89(7%) |
Nausea and/or vomiting | Binary | 164(9%) | 113(10%) |
Palpitations | Binary | 226(12%) | 164(14%) |
Other symptoms | Binary | 873(49%) | 618(54%) |
ECG Interpretation | |||
ST elevation | \(\{0,1\}^{11}\) | 329(18%) | 170(15%) |
ST depression | \(\{0,1\}^{11}\) | 500(28%) | 217(19%) |
T wave inversion | \(\{0,1\}^{11}\) | 252(14%) | 180(15%) |
